Speechtext

LICENSE.txt

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2022 SlashML
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

README.md

@@ -1,3 +1,66 @@
-# slashml-python-client
+# SlashML Python client
+[SlashML](https://www.slashml.com/)
+# Introduction
+## Overview 
+This is a Python client (SDK) for SlashML. It allows the user to use the apps available and active in [SlashML-Dashboard](https://www.slashml.com/dashboard). 
+The apps can be together in the same code. For example, if the user wants to transcribe an audio file and get a summary, they can call speechtotext followed by summarization.
 
-to be updated to guide users
+State-of-the-art AI models from several service providers are available. At SlashML, we also do the benchmarking on these models for the user. This will give them an idea of the best service provider for their application. For the full list of the available models thru SlashML, click [here](##availlable-service-providers)
+
+
+
+to lets you run transcription jobs from your Python code or Jupyter notebook. The transcription can be done with three lines of code
+```
+import slashml
+
+speect_to_text = slashml.SpeechToText()
+transcribe_id= speect_to_text.transcribe(audio_url,service_provider="aws")
+status=speect_to_text.status(transcribe_id,service_provider=service_provider)
+
+```
+
+## Set up and usage
+There is a daily limit (throttling) on the number of calls the user performs,  the code runs without specifying a token (API key),  the throttling kicks in and prevents new jobs after exceeding 10 calls per minute. 
+
+If the user intends on using the service more frequently, it is recommended to generate an token or API key from [SlashML-Dashboard](https://www.slashml.com/dashboard). This way, the throttling limit will increase.
+
+Sign up and Grab your token from [SlashML-Dashboard](https://www.slashml.com/dashboard)>settings> new api key and authenticate by setting it as an environment variable (or when you initialize the service, see Quickstart tutorial):
+
+In your terminal
+```
+export SLASHML_API_KEY=[token]
+```
+or including it in your python code as follows:
+```
+
+import os
+os.environ["SLASHML_API_KEY"] = "slashml_api_token"
+
+```
+
+If the user preferes using the API calls directly, the documentation is available [here](https://www.slashml.com/dashboard).
+## Availlable service providers
+
+### Speech-to-text
+
+AssemblyAI
+AWS
+Whisper (OpenAI)
+
+### Summarization
+hugging-face summarization based on Meta... include links
+# Quickstart tutorial 
+
+## Introduction
+
+### Start with initializing the service 
+
+### Specify your service provider
+
+In this step, benchmarking will help you decide which service provider is the best for you. 
+For speech-to-text: "assembly", "aws", "whisper"
+For summarization: "hugging-face", "openai"
+
+
+et voilà, Next steps:
+- pip install slashml

src/LICENSE.txt

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2022 SlashML
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

src/requirements.txt

@@ -0,0 +1 @@
+requests==2.28.1

src/slashml.py

@@ -0,0 +1,85 @@
+import requests
+import os
+from pathlib import Path
+import json
+
+class SpeechToText:
+
+    SLASHML_BASE_URL = 'https://api.slashml.com/speech-to-text/v1'
+    SLASHML_UPLOAD_URL = SLASHML_BASE_URL+'/upload/'
+    SLASHML_TRANSCRIPT_URL = SLASHML_BASE_URL+'/transcribe/'
+    SLASHML_STATUS_URL = SLASHML_BASE_URL+'/transcription'
+    SLASHML_TRANSCRIPT_STATUS_URL = lambda self,id: f"{SpeechToText.SLASHML_STATUS_URL}/{id}/"
+
+    HEADERS:dict = {}
+    ## add the api key to sys path envs 
+    def __init__(self,API_KEY: str = None):
+        self.API_KEY=None
+        if API_KEY:
+            token="Token {0}".format(API_KEY)
+            self.HEADERS = {'authorization': token}
+           # print("Auth with "+API_KEY+"\nMake sure this matches your API Key generated in the dashboard settings")
+        elif os.environ.get('SLASHML_API_KEY'):
+            key_env =  os.environ.get('SLASHML_API_KEY')
+            token="Token {0}".format(key_env)
+            self.HEADERS = {'authorization': token}
+           # print("Auth with environment variable SLASHML_API_KEY")
+        else:
+            self.HEADERS=None 
+            print("No Auth, there are certain limites to the usage")
+
+    def upload_audio(self, file_location:str,header=None):
+        headers = self.HEADERS
+        files=[
+        ('audio',('test_audio.mp3',open(file_location,'rb'),'audio/mpeg'))
+        ]
+        response = requests.post(self.SLASHML_UPLOAD_URL,
+                                headers=headers,files=files)
+        return  response.json()["upload_url"]
+
+    def transcribe(self,upload_url:str,  service_provider: str,header=None ):
+
+        transcript_request = {'audio_url': upload_url}
+        headers = self.HEADERS
+        payload = {
+        "uploaded_audio_url": upload_url,
+        "service_provider": service_provider
+        }
+        response = requests.post(self.SLASHML_TRANSCRIPT_URL, headers=headers, data=payload)
+        # Check the status code of the response
+        if response.status_code == 200:
+            return response.json()["id"]
+
+        elif response.status_code == 429:
+            return "THROTTLED"
+
+        elif response.status_code == 404:
+            return "NOT FOUND"
+
+        elif response.status_code == 500:
+            return "SERVER ERROR"
+        else:
+            return "UNKNOWN ERROR"
+
+    def status(self,job_id:str, service_provider: str ,header=None):
+        headers = self.HEADERS
+        payload = {
+        "service_provider": service_provider
+        }
+
+        response = requests.get(self.SLASHML_TRANSCRIPT_STATUS_URL(job_id) , headers=headers, data=payload)
+
+        # Check the status code of the response
+        if response.status_code == 200:
+            return response.json()["transcription_data"]["transcription"]
+
+        elif response.status_code == 429:
+            return "THROTTLED"
+
+        elif response.status_code == 404:
+            return "NOT FOUND"
+
+        elif response.status_code == 500:
+            return "SERVER ERROR"
+        else:
+            return "UNKNOWN ERROR"

src/test_speechtotext_1.py

@@ -0,0 +1,22 @@
+import slashml
+import os 
+
+#API KEY (optional)
+
+# set environment path
+import os
+os.environ["SLASHML_API_KEY"] = "0d91bfede9c5c9de6ff1d5610ef71c3b6d5be9ee"
+# Initialize SlashML 
+speect_to_text = slashml.SpeechToText()
+# optional local file to upload, if not an accessible url
+file_location="/Users/JJneid/Desktop/SlashMl/Benchmarking/podcast1/podcast1_long_trim.mp3"
+# If your audio files aren't accessible via a URL already, you can upload your audio file using this API
+upload_url= speect_to_text.upload_audio(file_location)
+# choose your service provider: "assembly", "aws", "whisper"
+service_provider="aws"
+# transcribe
+transcribe_id= speect_to_text.transcribe(upload_url,service_provider)
+print(transcribe_id)
+
+status=speect_to_text.status(transcribe_id,service_provider=service_provider)
+print(status)